Numerical inversion of probability generating functions (Q1200798)

The authors provide an algorithm to approximate the discrete probability distribution \((p_ k)_{k\geq 0}\) when only the probability generating function \(G(z)=\sum_{k=0}^ \infty q_ k z^ k\), \(z\in\mathbb{C}\), \(| z|<1\), is known. The algorithm is based on numerical inversion of the probability generating function using the Fourier series method. Bounds for the approximation are given. As an example the distribution of the number of customers served in a busy period of an \(M/M/1\)-queue is considered.